Figure 4

Network analysis applying the ‘PANDA’ method to Hhip ^+/- mice vs. Hhip ^+/+ mice treated with either room air (Air) or cigarette smoke (CS) for 6 months. (A) Diagram to illustrate building networks using PANDA to combine transcription factor motif and physical protein interaction data with gene expression data from each group using a jack-knife subsampling method. (B) A volcano plot illustrating the selection of subnetworks by choosing high-probability edges specific to either the Hhip ^+/--CS (red) or Hhip ^+/+-CS (blue) model. Red lines illustrate the thresholds used to select these edges, which were based on both the difference in average edge-weight (vertical lines) as well as P value significance (horizontal line). (C) Enriched KEGG Pathways (false discovery rate (FDR) <0.01) associated with at least one of eight distinct subnetworks, defined by four different comparisons (each column). FDR statistical significance is shown as a color with red colors representing more significant enrichment in subnetwork 1 (that is, Hhip ^+/- -CS in column 3) and blue colors representing more significant enrichment in subnetwork 2 (that is, Hhip ^+/+-CS in column 3). Blue stars indicate lymphocyte activation pathway.